Information Advantage Inc reckons that the next version of its DecisionSuite data warehouse offering is the first to address the scalability issues associated with customer-centric data warehouses, as opposed to financial or product-centric ones. The Minneapolis, Minnesota company can also cite MasterCard International Inc among its customers and is already rolling out DecisionSuite 3.5 in a 1.2Tb configuration supporting 22,000 member banks (CI No 2,853). The new version includes the company’ s WebOLAP front-end to DecisionSuite, which trusts the World Wide Web to enable users to analyze the most up-to-date version of the data. Naturally, Information Advantage was at pains to emphasize the multi-layer security. Each user has a security p rofile on the server at personal, workgroup and enterprise levels, accessed by their encrypted password. There are also layers of database security, whereby users can have collaborative capabilities but also lock-outs. Each Web page has a token atta ched to it that retains the user’s security ID so they do not have to log on again if they want to go back and look at the page in the same session. Customer-centric data warehouses are the most demanding in terms of the size of the database. They t end to be used in industries with millions of customers, such as retail, insurance, telecommunications and finance. Whereas financial data warehouses typically range from 5Gb to 20Gb, and product ones from 50Gb to 500Gb – the company’s previous area of focus – customer data warehouses can be multi-Terabyte. This provides a set of problems that Information Advantage believes it has met. For instance, customer analysis databases tend to have large numbers of users. And if they are all making SQL queries on it using temporary tables, it does not take long to slow the whole thing down significantly. Therefore, DecisionSuite aims to get around this by eliminating the need for temporary tables by optimizing processing between the database and Decision-Suite server through what the company calls join, aggregate and data filter optimization. Join optimization entails switching on the fly between logical joins – those in the DecisionSuite analytical processing engine – and physical joins – those done in the actual customer database. For example, the product data join is done in the DecisionSuite server, and the customer join and segmentation into groups would be done in the physical database. Aggregate optimization involves processing raw data, such as unit sales in the database, and calculated data, such as the percentage market share in the DecisionSuite server, again eliminating the need for temporary calculation tables and improving performance, according to the company. Data filter optimization Lastly, data filter optimization auto-mates where filt-ering takes place, again with the raw data filtering in the database and the processed data in the analytical processing engine on the Decision-Suite server. DecisionSuite also has native SQL su pport that tunes the SQL statements to the particular vendor’s relational database and splits large statements into multiple queries. Decision-Suite 3.5 comes with four client applications, Info-Alert for light users; NewsLine, which enables users t o modify reports; Workbench, an administrator’s client and Analysis, for power users to publish their own reports and develop their own intelligent agents. Available from this month for Windows clients supporting Object Linking & Embedding, Messagin g Applications Programming Interface, Dynamic Data Exchange and extendible using Visual Basic, C++ or PowerBuilder, and a wide variety of Unix servers, it also supports DB2/6000, Informix, Oracle, Red Brick, Sybase and Tandem NonStop SQL databases. Decision-Suite server with WebOLAP goes from $30,000 and the Info-Alert, Newsline, Analy-is and Workbench clients are $45, $150, $900 and $3,000 respectively.
